Numerical Simulation

The geometry and the loads on the specimen used in the component tests were modeled with the ANSYS FEM software[24] in a numerical simulation. In this sim­ulation, the glass and the adhesive joint were modeled with volume elements

TABLE 5—Material constants.

Material

E (MPa)

l (-)

Glass

70,000

0.23

Adhesive

234

0.42

(SOLID 186, 3D 20-node structural solid). The plastic setting block was omitted from the model. For reasons of symmetry, calculation of half the glass structure was adequate. The material constants used in the numerical analysis are shown in Table 5. It was assumed that strains remain small and lie below the yield point of the adhesive, so a linear-elastic material behavior could be assumed. The Young’s modulus of the adhesive was determined according to Ref [18] using the average values of the stress-strain relation (Fig. 6) at room tempera­ture. The model was first applied for planning the component tests to determine the positions of maximum tensile stress in the glass. It was at these positions that strain gauges for monitoring the tensile stresses were attached for the tests. In a second step, the results of the numerical simulation (Figs. 14-16) were compared with the values measured during the component tests.

The maximum tensile stresses were found to occur in each case at the tran­sition from the frame beam or frame post to the adhesive. This is where only two glass plies are involved in transferring the load. Table 6 shows the calcu­lated and measured stress and deformation values at the critical points for a

FIG. 14—Results of the numerical analysis with a load of 80 kN – overall deformations.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

FIG. 15—Results of the numerical analysis with a load of 80 kN – stresses in x direction in glass cantilever.

FIG. 16—Results of the numerical analysis with a load of 80 kN – stresses in y direction in glass post.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

Rochester Institute Of Technology pursuant to License Agreement. No further reproductions authorized.

TABLE 6—Comparison of calculated and measured values at a load of 80 kN.

Strain gauge position

1

2

5

6

Specimen component

Glass

Glass

Glass

Glass

Deflection at

Load

stress

stress

stress

stress

end of cantilever

(kN)

rg (MPa)

rg (MPa)

rg (MPa)

rg (MPa)

fy (mm)

SC 1

80.0

138.9

139.4

110.9

108.9

4.0

SC 2

80.0

136.5

108.7

109.7

3.9

SC 3

80.0

136.4

141.8

108.7

113.8

4.1

SC 5

80.0

137.4

136.9

3.9

SC 6

80.0

141.0

128.4

113.3

104.8

4.3

SC 7

80.0

130.3

142.8

104.8

116.1

3.8

SC 8

80.0

136.1

140.0

112.3

96.8

3.8

SC 9

80.0

144.0

132.9

3.9

SC 10

80.0

124.5

148.5

3.9

Measured value (mean)

136.1

138.8

109.8

108.3

4.0

FE calculated

135.2

135.2

108.2

108.2

5.1

Deviation

-0.7 %

-2.6 %

-1.5 %

-0.1 %

+29.0 %

load of 80 kN. There was good agreement between the calculated values and the mean values obtained from experimental studies. The model predicts the stresses in the glass with very good accuracy. The deformations are slightly overestimated by the numerical calculations but are, therefore, on the safe side. This may result from the identification of the material properties on the bulk material. Assuming a linear-elastic material behavior for modeling the adhesive is, therefore, suitable for the numerical analysis of the global load-bearing sys­tem. As a next step, the numerical model may be utilized for a sensitivity analy­sis focusing on the structural behavior at different temperature levels.

Conclusions

Adhesives enable individual glass elements to be combined to form transparent load-bearing structures. Previous research work and a construction project that has already been realized enabled the derivation of key requirement criteria that must be met by an adhesive suitable for a glued frame corner. Preliminary studies enabled an adhesive to be determined that exhibits improved working properties compared with the composition used in the past and at the same time still satisfies the structural requirements. The load-carrying capacity of the glued connection at room temperature was confirmed by tests on components. A setting block provided for the failure case of the adhesive joint was shown to have no effect on the global load-bearing behavior of the frame corner.

Acknowledgments

The authors would like to thank the German Federal Ministry of Economics and Technology for supporting this research project. Our thanks are also

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

Rochester Institute Of Technology pursuant to License Agreement. No further reproductions authorized.

extended to the research project partners GSK—Glas Statik Konstruktion, First

Glas, and Bayer Glasbau in Germany, as well as 3M, DELO Industrial Adhe­sives, and Henkel, Huntsman, and Panacol for supplying the adhesives.